Adding out-of-bag errors and convergence monitoring

[bgall]

Two enhancements would be very useful:

1. Add an argument to return the out-of-bag (OOB) prediction error (on a variable-by variable basis), a la missForest. If argument is T, could return a list with data in one element and error in others. Preferably variablewise. Right now only the average is returned.

2. Some way of monitoring convergence like mice's plots of the chained values.

[mayer79]

Thanks for these hints!

1. I think the first point is captured by the returnOOB argument. It attaches the attribute oob to the resulting dataset with the OOB errors of the final round. I am aware this interface is not ideal. Nowadays, I'd definitively return a list with class "missRanger" and a couple of methods, like e.g. in my outForest outlier detection package. Maybe I can go into the direction you are proposing and go for an argument return_data_only = TRUE for the current behaviour and for a nice "missRanger" object otherwise. I will put this on my "Covid summer holiday" to do list!

2. Good idea. I will consider adding this feature to the next release.

[mayer79]

Finally solved via #55